Our long-term objective is to continue studies of color vision in 2- to 3-month-old infants. The properties of the infant's first (photoreceptor) stage are already known to be similar to those of adults. We now plan to study the second ("early opponent") stage of color coding, and its transition to higher level codes. Our working hypothesis can be called the high level immaturity hypothesis: that infants will show an adultlike early opponent code, but will fail to show the signatures of higher level mechanisms such as a brightness code and higher order color codes. The data will also address several other current developmental hypotheses, including uniform vs. differential loss, adult-like vs. juvenile mechanisms, and rod dominance. Series 1: The luminance channel. The first and simplest second-stage mechanism is the luminance channel. We already know that infants have a luminance channel. This series of experiments will further define and quantify its properties, including the magnitudes of individual differences among infants. Other data will address the rod dominance hypothesis and the transition from a luminance to a (higher level) brightness code. Series 2: The r/g channel. In this series we add the "red/green" (r/g) dimension of color with studies in the L-cone, M-cone plane of color space. Discrimination ellipses will be used to get an atheoretical estimate of the color code (privileged axes) in infants. Cross-masking will be used to define further whether signals on some pairs of axes of color space interact while others do not. The results will address the hypotheses of uniform vs. differential loss and adult-like vs. juvenile mechanisms for the luminance and r/g dimensions. Series 3: The tritan channel. In this series we add the early "blue/yellow" (tritan) dimension by using stimuli centered on the "white" point in 3-D color space. Discrimination ellipses and cross-masking experiments will be executed as in Series 2. In addition, we will use a cross-masking paradigm (sectored noise) that in adults reveals a recoding to a third level (higher-order) color code. The results will particularly address the hypotheses of uniform vs. differential loss and adult-like vs. juvenile mechanisms in the isoluminant plane.